The Power of Behavioral Data | Verikai

The Power of Behavioral Data

It is estimated that, on average, we make 35,000 decisions every day. Each time we make a decision and it’s recorded in the cybersphere, we are contributing to not only the vast universe of big data, but also to our own data profiles that reflect our individual choices.

I was destined to work in the “family business” of insurance. But, as is the typical story of headstrong children, I made my own path to getting here. Let me explain. 

My mother served as a successful actuary for a major insurance company in the auto industry. She was one of the key stakeholders involved in uncovering the insights that eventually led to the development of the Name Your Own Price tool. It is no doubt that I get my curiosity for data from her. 

Like most Indian mothers, she wanted me to become a doctor. And so I did. But my passion for data and statistics was always stronger. While I am still a practicing physician, I turned my attention to entrepreneurship many years ago. In my third company, Verikai, my life has come full circle. We are using data and machine learning to improve the world of healthcare underwriting.

It may not be exactly what my mom had in mind, but the experience I gained from my clinical days led me to where I am today and was an important step in uncovering the power of behavioral data. Thanks, Mom!

(And, if you’re wondering, she still tells other parents that I’m a doctor, so it’s very clear to others that she raised a respectable son.)

What is behavioral data? 

It is estimated that, on average, we make 35,000 decisions every day. Cross reference this with the staggering number of data points recorded every day — 2.5 quintillion bytes — and it’s easy to recognize the vast potential of this untapped territory. Each time we make a decision and it’s recorded in the cybersphere, we are contributing to not only the vast universe of big data, but also to our own data profiles that reflect our individual choices. 

Therein lies the definition of behavioral data — the information gathered on an individual as a result of their actions. And in insurance, more specifically, the actions we take as a result of being presented with risk. Understanding human behavioral patterns helps us understand why individuals make the choices they do, and what the likely impact of those choices will be on the financial outcomes that are meaningful in their lives at that time. 

What will a patient do when they leave the doctor’s office with new information about their health risks? What will they do in three months? Does that answer change if their decision doesn’t have an immediate impact? What will these decisions cost them and their insurer? The answer lies in their data. And (spoiler alert), this data is highly predictive of action or in-action with respect to his or her own health.

Genetics and demographics versus behavior 

While genetics is one of the strongest predictors of disease, behavior tells us what is arguably more important to impacting the outcome. Not all medical conditions are preventable or even predictable, but a large and growing percentage of medical expenditures are focused on conditions that are linked to lifestyle choices (cancer, heart disease, diabetes, cirrhosis, etc.) And, unlike genetics, these lifestyle choices can be changed. This is why carriers are spending so much money in population management and behavior modification.

While demographics are the most commonly sought after data point in understanding human behavior across all industries (the insurance industry being no exception), there are pervasive examples that demonstrate the dangers of overgeneralization. Consider these:

  • Not every married woman between the ages of 25 to 34 is planning to give birth. 
  • Not every teenage driver will drive recklessly and get into an accident. 
  • Not every homeowner with a pool will have someone drown on their property. 
  • Not every roofing contractor will fall off a roof. 

Yet, the insurance industry sets pricing based on these generalizations, which results in a financial penalization of the entire group because they cannot discern the difference between good actors and bad actors. To remain profitable, insurance carriers need everyone to pay more in order to compensate for the percentage of people who will actually cost them more to cover.

Seem unfair? That’s because it is. 

The reality is that some of us take better care of ourselves than others. And while much about our health is beyond our control (for example, our genetics), we make medical and lifestyle choices every day that contribute to our overall wellbeing. For example, individuals with markers of higher socio-economic status (they are well educated, live in nice houses, travel, have lots of friends), tend to live healthier lives due to their access to an ecosystem of healthier lifestyle options. 

A well documented example of this concept, known as the social determinants of health, occurs in patients with Type 2 diabetes. Those with a more positive socio-ecological and socioeconomic status are more likely to take steps toward living a healthier lifestyle, which ultimately leads to a reduction in the complications associated with their disease. But again, this is a generalization. Not everyone will fit this stereotype, which is why it’s important to dig deeper to understand behavior at the individual level. Going beyond generalization is the critical next step in the evolution of underwriting.

The importance of behavior to the insurance business

When determining the level of risk the insured will bring, it’s not enough to understand the potential diagnosis of a patient, but equally important is whether or not they will adhere to medical recommendations and best practices. Will they take their medication? Will they change their lifestyle? Will the person use emergency services to cope with his or her condition or function primarily as an outpatient through preventative services? The answers to all of these questions impact the overall long term healthcare cost of the patient.

In health insurance, the typical pricing model is built on the law of averages. Some patients will comply with the doctor’s recommendations, and some won’t. This allows us to be fairly accurate in our assessment of risk and potential healthcare costs, but it is a model that is still inaccurately based on generalizations and unfairly penalizes those who do have a track record of making healthy decisions. 

For insurance carriers, this used to be good enough. Conversative pricing reduced their own level of risk and protected them from catastrophic claims. However, in today’s oversaturated health insurance market (currently more than 900 health insurance companies operate in the United States), a carrier’s ability to grow is tied to its ability to lower prices without inheriting a high probability of risk. This conundrum is exacerbated by increasing competition with other insurance carriers that offer self-insured health products to employers who are looking to reduce healthcare costs. It means that carriers have to start rewarding good behavior and only punishing bad behavior, to protect their profit.

Verikai’s Capture for Health allows us to assess risk based on the behavioral choices of individuals within a group’s census. Layered on top of additional data sources, this becomes a powerful tool: Who is most likely to submit a health insurance claim? How frequently will that person submit a claim? What will the total cost of their claims be? How much does this total cost deviate from the expectation, for both the individual and the employer group? 

Knowing these things in advance brings with it a level of confidence that has never existed in the health insurance underwriting world – one that allows you to be more accurate, more competitive and more profitable. 



Leave a Reply

Your email address will not be published. Required fields are marked *